Message Passing Programming. Designing MPI Applications
|
|
- Frank Harmon
- 5 years ago
- Views:
Transcription
1 Message Passing Programming Designing MPI Applications
2 Reusing this material This work is licensed under a Creative Commons Attribution- NonCommercial-ShareAlike 4.0 International License. This means you are free to copy and redistribute the material and adapt and build on the material under the following terms: You must give appropriate credit, provide a link to the license and indicate if changes were made. If you adapt or build on the material you must distribute your work under the same license as the original. Note that this presentation contains images owned by others. Please seek their permission before reusing these images. 2
3 Overview Lecture will cover - MPI portability - maintenance of serial code - general design - debugging - verification 3
4 MPI Portability Potential deadlock you may be assuming that MPI_Send is asynchronous it often is buffered for small messages but threshold can vary with implementation a correct code should run if you replace all MPI_Send calls with MPI_Ssend Buffer space cannot assume that there will be space for MPI_Bsend default buffer space is often zero! be sure to use MPI_Buffer_Attach some advice in MPI standard regarding required size 4
5 Data Sizes Be careful of data sizes or layout - use runtime enquiry functions for Fortran types - be careful of compiler-dependent padding for structures Changing precision - when changing from, say, float to double, must change all the MPI types from MPI_FLOAT to MPI_DOUBLE as well Easiest to achieve with an include file - e.g. every routine includes precision.h 5
6 Changing Precision: C Define a header file called, e.g. precision.h - typedef float RealNumber - #define MPI_REALNUMBER MPI_FLOAT Include in every function - #include precision.h RealNumber x; - MPI_Routine(&x, MPI_REALNUMBER,...); Global change of precision now easy - edit 2 lines in one file: float->double, MPI_FLOAT->MPI_DOUBLE 6
7 Changing Precision: Fortran Define a module called, e.g., precision - integer, parameter :: REALNUMBER=kind(1.0e0) - integer, parameter :: MPI_REALNUMBER = MPI_REAL Use in every subroutine - use precision REAL(kind=REALNUMBER):: x - call MPI_ROUTINE(x, MPI_REALNUMBER,...) Global change of precision now easy - change 1.0e0 -> 1.0d0, MPI_REAL-> MPI_DOUBLE_PRECISION 7
8 Testing Portability Run on more than one machine - assuming the implementations are different - many parallel clusters will use the same open-source MPI e.g. OpenMPI or MPICH2 running on two different mid-sized machines may not be a good test More than one implementation on same machine - e.g. run using both MPICH2 and OpenMPI on your laptop - very useful test, and can give interesting performance numbers More than one compiler - user@cluster$ module switch mpich2-pgi mpich2-gcc 8
9 Serial Code Adding MPI can destroy a code - would like to maintain a serial version - i.e. can compile and run identical code without an MPI library - not simply running MPI code with P=1! Need to separate off communications routines - put them all in a separate file - provide a dummy library for the serial code - no explicit reference to MPI in main code 9
10 Example: Initialisation! parallel routine subroutine par_begin(size, procid) implicit none integer :: size, procid include "mpif.h" call mpi_init(ierr) call mpi_comm_size(mpi_comm_world, size, ierr) call mpi_comm_rank(mpi_comm_world, procid, ierr) procid = procid + 1 end subroutine par_begin! dummy routine for serial machine subroutine par_begin(size, procid) implicit none integer :: size, procid size = 1 procid = 1 end subroutine par_begin 10
11 Example: Global Sum! parallel routine subroutine par_dsum(dval) implicit none include "mpif.h" double precision :: dval, dtmp call mpi_allreduce(dval, dtmp, 1, MPI_DOUBLE_PRECISION, & MPI_SUM, comm, ierr) dval = dtmp end subroutine par_dsum! dummy routine for serial machine subroutine par_dsum(dval) implicit none double precision dval end subroutine par_dsum 11
12 Example Makefile SEQSRC= \ demparams.f90 demrand.f90 demcoord.f90 demhalo.f90 \ demforce.f90 demlink.f90 demcell.f90 dempos.f90 demons.f90 MPISRC= \ demparallel.f90 \ demcomms.f90 FAKESRC= \ demfakepar.f90 \ demfakecomms.f90 #PARSRC=$(FAKESRC) PARSRC=$(MPISRC) 12
13 Advantages of Comms Library Can compile serial program from same source - makes parallel code more readable Enables code to be ported to other libraries - more efficient but less versatile routines may exist - e.g. Cray-specific SHMEM library - can even choose to only port a subset of the routines Library can be optimised for different MPIs - e.g. choose the fastest send (Ssend, Send, Bsend?) 13
14 Design Separate the communications into a library Make parallel code similar as possible to serial - e.g. use of halos in case study - could use the same update routine in serial and parallel serial: update(new, old, M, N ); parallel: update(new, old, MP, NP); - may have a large impact on the design of your serial code Don t try and be too clever - don t agonise whether one more halo swap is really necessary - just do it for the sake of robustness 14
15 General Considerations Compute everything everywhere - e.g. use routines such as Allreduce - perhaps the value only really needs to be know on the master but using Allreduce makes things simpler no serious performance implications Often easiest to make P a compile-time constant - may not seem elegant but can make coding much easier e.g. definition of array bounds - put definition in an include file - a clever Makefile can reduce the need for recompilation only recompile routines that define arrays rather than just use them pass array bounds as arguments to all other routines 15
16 Debugging Parallel debugging can be hard Don t assume it s a parallel bug! - run the serial code first - then the parallel code with P=1 - then on a small number of processes Writing output to separate files can be useful - e.g. log.00, log.01, log.02,. for ranks 0, 1, 2,... - need some way easily to switch this on and off Some parallel debuggers exist - Totalview is the leader across all largest platforms - Allinea DDT is becoming more common across the board 16
17 General Debugging People seem to write programs DELIBERATELY to make them impossible to debug! - my favourite: the silent program - my program doesn t work $ mprun np 6./program.exe $ SEGV core dumped - where did this crash? - did it run for 1 second? 1 hour? in a batch job this may not be obvious - did it even start at all? Why don t people write to the screen!!! 17
18 Program should output like this $ mprun np 6./program.exe Program running on 6 processes Reading input file input.dat done Broadcasting data done rank 0: x = 3 rank 1: x = 5 etc etc Starting iterative loop iteration 100 iteration 200 finished after 236 iterations writing output file output.dat done rank 0: finished rank 1: finished Program finished 18
19 Typical mistakes Don t write raw numbers to the screen! - what does this mean? $ mprun np 6./program.exe programmer has written $ printf( %d %d %f\n, rank, j, x); $ write(*,*) rank, j, x Takes an extra 5 seconds to type: $ printf( rank, j, x: %d %d %f\n, rank, j, x); $ write(*,*) rank, j, x:, rank, j, x - and will save you HOURS of debugging time Why oh why do people write raw numbers?!?! 19
20 Debugging walkthrough My case study code gives the wrong answer Stages: - read data in - distribute to processes - update many times requiring halo swaps - collect data back - write data out Final stage shows the error - but where did it first go wrong? 20
21 Where is it going wrong? On input? On distribute? On update? - on halo swaps? - on left/right swaps? - on up/down swaps? On collection? On output? All these can be checked with simple tests 21
22 Common mistake I changed something - and it now works (but I don t know why) All is OK! No! - there is a bug - you MUST find it - if not, it will come back later to bite you HARD Debugging is an experimental science 22
23 Verification: Is My Code Working? Should the output be identical for any P? - very hard to accomplish in practice due to rounding errors may have to look hard to see differences in the last few digits - typically, results vary slightly with number of processes - need some way of quantifying the differences from serial code - and some definition of acceptable What about the same code for fixed P? - identical output for two runs on same number of processes? - should be achievable with some care not in specific cases like dynamic task farms possible problems with global sums MPI doesn t require reproducibility, but most implementations are - without this, debugging is almost impossible 23
24 Parallelisation Some parallel approaches may be simple - but not necessarily optimal for performance - case study example is very simple due to 1D decomposition but not particularly efficient for large P - often need to consider what is the realistic range of P Some people write incredibly complicated code - step back and ask: what do I actually want to do? - is there an existing MPI routine or collective communication? - should I reconsider my approach if it prohibits me from using existing routines, even if it is not quite so efficient? 24
25 Optimisation Keep running your code - on a number of input data sets - with a range of MPI processes If scaling is poor - find out what parallel routines are the bottlenecks - again, much easier with a separate comms library If performance is poor - work on the serial code - return to parallel issues later on 25
26 Conclusions Run on a variety of machines Keep it simple Maintain a serial version Don t assume all bugs are parallel bugs Find a debugger you like (good luck to you) 26
Message Passing Programming. Modes, Tags and Communicators
Message Passing Programming Modes, Tags and Communicators Reusing this material This work is licensed under a Creative Commons Attribution- NonCommercial-ShareAlike 4.0 International License. http://creativecommons.org/licenses/by-nc-sa/4.0/deed.en_us
More informationParallel Programming. Libraries and Implementations
Parallel Programming Libraries and Implementations Reusing this material This work is licensed under a Creative Commons Attribution- NonCommercial-ShareAlike 4.0 International License. http://creativecommons.org/licenses/by-nc-sa/4.0/deed.en_us
More informationMessage-Passing Programming with MPI. Message-Passing Concepts
Message-Passing Programming with MPI Message-Passing Concepts Reusing this material This work is licensed under a Creative Commons Attribution- NonCommercial-ShareAlike 4.0 International License. http://creativecommons.org/licenses/by-nc-sa/4.0/deed.en_us
More informationMessage Passing Programming. Introduction to MPI
Message Passing Programming Introduction to MPI Reusing this material This work is licensed under a Creative Commons Attribution- NonCommercial-ShareAlike 4.0 International License. http://creativecommons.org/licenses/by-nc-sa/4.0/deed.en_us
More informationParallel Programming Libraries and implementations
Parallel Programming Libraries and implementations Partners Funding Reusing this material This work is licensed under a Creative Commons Attribution- NonCommercial-ShareAlike 4.0 International License.
More informationCSE. Parallel Algorithms on a cluster of PCs. Ian Bush. Daresbury Laboratory (With thanks to Lorna Smith and Mark Bull at EPCC)
Parallel Algorithms on a cluster of PCs Ian Bush Daresbury Laboratory I.J.Bush@dl.ac.uk (With thanks to Lorna Smith and Mark Bull at EPCC) Overview This lecture will cover General Message passing concepts
More informationReusing this material
Messages Reusing this material This work is licensed under a Creative Commons Attribution- NonCommercial-ShareAlike 4.0 International License. http://creativecommons.org/licenses/by-nc-sa/4.0/deed.en_us
More informationReusing this material
Derived Datatypes Reusing this material This work is licensed under a Creative Commons Attribution- NonCommercial-ShareAlike 4.0 International License. http://creativecommons.org/licenses/by-nc-sa/4.0/deed.en_us
More informationParallel Programming. Libraries and implementations
Parallel Programming Libraries and implementations Reusing this material This work is licensed under a Creative Commons Attribution- NonCommercial-ShareAlike 4.0 International License. http://creativecommons.org/licenses/by-nc-sa/4.0/deed.en_us
More informationFractals. Investigating task farms and load imbalance
Fractals Investigating task farms and load imbalance Reusing this material This work is licensed under a Creative Commons Attribution- NonCommercial-ShareAlike 4.0 International License. http://creativecommons.org/licenses/by-nc-sa/4.0/deed.en_us
More informationFractals exercise. Investigating task farms and load imbalance
Fractals exercise Investigating task farms and load imbalance Reusing this material This work is licensed under a Creative Commons Attribution- NonCommercial-ShareAlike 4.0 International License. http://creativecommons.org/licenses/by-nc-sa/4.0/deed.en_us
More informationAMath 483/583 Lecture 21
AMath 483/583 Lecture 21 Outline: Review MPI, reduce and bcast MPI send and receive Master Worker paradigm References: $UWHPSC/codes/mpi class notes: MPI section class notes: MPI section of bibliography
More informationMessage Passing Programming. Modes, Tags and Communicators
Message Passing Programming Modes, Tags and Communicators Overview Lecture will cover - explanation of MPI modes (Ssend, Bsend and Send) - meaning and use of message tags - rationale for MPI communicators
More informationMPI Casestudy: Parallel Image Processing
MPI Casestudy: Parallel Image Processing David Henty 1 Introduction The aim of this exercise is to write a complete MPI parallel program that does a very basic form of image processing. We will start by
More informationCFD exercise. Regular domain decomposition
CFD exercise Regular domain decomposition Reusing this material This work is licensed under a Creative Commons Attribution- NonCommercial-ShareAlike 4.0 International License. http://creativecommons.org/licenses/by-nc-sa/4.0/deed.en_us
More informationParallel Programming Using MPI
Parallel Programming Using MPI Short Course on HPC 15th February 2019 Aditya Krishna Swamy adityaks@iisc.ac.in SERC, Indian Institute of Science When Parallel Computing Helps? Want to speed up your calculation
More informationParallel Programming Using Basic MPI. Presented by Timothy H. Kaiser, Ph.D. San Diego Supercomputer Center
05 Parallel Programming Using Basic MPI Presented by Timothy H. Kaiser, Ph.D. San Diego Supercomputer Center Talk Overview Background on MPI Documentation Hello world in MPI Basic communications Simple
More informationNon-Blocking Communications
Non-Blocking Communications Reusing this material This work is licensed under a Creative Commons Attribution- NonCommercial-ShareAlike 4.0 International License. http://creativecommons.org/licenses/by-nc-sa/4.0/deed.en_us
More informationRecap of Parallelism & MPI
Recap of Parallelism & MPI Chris Brady Heather Ratcliffe The Angry Penguin, used under creative commons licence from Swantje Hess and Jannis Pohlmann. Warwick RSE 13/12/2017 Parallel programming Break
More informationSlides prepared by : Farzana Rahman 1
Introduction to MPI 1 Background on MPI MPI - Message Passing Interface Library standard defined by a committee of vendors, implementers, and parallel programmers Used to create parallel programs based
More informationProgramming with MPI
Programming with MPI p. 1/?? Programming with MPI One-sided Communication Nick Maclaren nmm1@cam.ac.uk October 2010 Programming with MPI p. 2/?? What Is It? This corresponds to what is often called RDMA
More informationFurther MPI Programming. Paul Burton April 2015
Further MPI Programming Paul Burton April 2015 Blocking v Non-blocking communication Blocking communication - Call to MPI sending routine does not return until the send buffer (array) is safe to use again
More informationDebugging Applications Using Totalview
Debugging Applications Using Totalview Timothy H. Kaiser tkaiser@sdsc.edu Slides by:nicholas J Wright First things first.. There is no point in optimising a program that has a bug in it a really fast code
More informationPractical Introduction to Message-Passing Interface (MPI)
1 Practical Introduction to Message-Passing Interface (MPI) October 1st, 2015 By: Pier-Luc St-Onge Partners and Sponsors 2 Setup for the workshop 1. Get a user ID and password paper (provided in class):
More informationBuilding Library Components That Can Use Any MPI Implementation
Building Library Components That Can Use Any MPI Implementation William Gropp Mathematics and Computer Science Division Argonne National Laboratory Argonne, IL gropp@mcs.anl.gov http://www.mcs.anl.gov/~gropp
More informationIntroduction to parallel computing with MPI
Introduction to parallel computing with MPI Sergiy Bubin Department of Physics Nazarbayev University Distributed Memory Environment image credit: LLNL Hybrid Memory Environment Most modern clusters and
More informationParallel Programming with MPI: Day 1
Parallel Programming with MPI: Day 1 Science & Technology Support High Performance Computing Ohio Supercomputer Center 1224 Kinnear Road Columbus, OH 43212-1163 1 Table of Contents Brief History of MPI
More informationAssignment 3 MPI Tutorial Compiling and Executing MPI programs
Assignment 3 MPI Tutorial Compiling and Executing MPI programs B. Wilkinson: Modification date: February 11, 2016. This assignment is a tutorial to learn how to execute MPI programs and explore their characteristics.
More informationMPI: The Message-Passing Interface. Most of this discussion is from [1] and [2].
MPI: The Message-Passing Interface Most of this discussion is from [1] and [2]. What Is MPI? The Message-Passing Interface (MPI) is a standard for expressing distributed parallelism via message passing.
More informationIntroduction to MPI. May 20, Daniel J. Bodony Department of Aerospace Engineering University of Illinois at Urbana-Champaign
Introduction to MPI May 20, 2013 Daniel J. Bodony Department of Aerospace Engineering University of Illinois at Urbana-Champaign Top500.org PERFORMANCE DEVELOPMENT 1 Eflop/s 162 Pflop/s PROJECTED 100 Pflop/s
More informationProgramming with MPI
Programming with MPI p. 1/?? Programming with MPI Miscellaneous Guidelines Nick Maclaren Computing Service nmm1@cam.ac.uk, ext. 34761 March 2010 Programming with MPI p. 2/?? Summary This is a miscellaneous
More informationMessage-Passing and MPI Programming
Message-Passing and MPI Programming 2.1 Transfer Procedures Datatypes and Collectives N.M. Maclaren Computing Service nmm1@cam.ac.uk ext. 34761 July 2010 These are the procedures that actually transfer
More informationIntroduction to Parallel Programming. Martin Čuma Center for High Performance Computing University of Utah
Introduction to Parallel Programming Martin Čuma Center for High Performance Computing University of Utah mcuma@chpc.utah.edu Overview Types of parallel computers. Parallel programming options. How to
More informationMessage Passing Interface
MPSoC Architectures MPI Alberto Bosio, Associate Professor UM Microelectronic Departement bosio@lirmm.fr Message Passing Interface API for distributed-memory programming parallel code that runs across
More informationAn introduction to MPI
An introduction to MPI C MPI is a Library for Message-Passing Not built in to compiler Function calls that can be made from any compiler, many languages Just link to it Wrappers: mpicc, mpif77 Fortran
More informationThe Message Passing Interface (MPI) TMA4280 Introduction to Supercomputing
The Message Passing Interface (MPI) TMA4280 Introduction to Supercomputing NTNU, IMF January 16. 2017 1 Parallelism Decompose the execution into several tasks according to the work to be done: Function/Task
More informationReusing this material
Modules Reusing this material This work is licensed under a Creative Commons Attribution- NonCommercial-ShareAlike 4.0 International License. http://creativecommons.org/licenses/by-ncsa/4.0/deed.en_us
More informationHolland Computing Center Kickstart MPI Intro
Holland Computing Center Kickstart 2016 MPI Intro Message Passing Interface (MPI) MPI is a specification for message passing library that is standardized by MPI Forum Multiple vendor-specific implementations:
More informationSINGLE-SIDED PGAS COMMUNICATIONS LIBRARIES. Basic usage of OpenSHMEM
SINGLE-SIDED PGAS COMMUNICATIONS LIBRARIES Basic usage of OpenSHMEM 2 Outline Concept and Motivation Remote Read and Write Synchronisation Implementations OpenSHMEM Summary 3 Philosophy of the talks In
More informationIntroduction to parallel computing concepts and technics
Introduction to parallel computing concepts and technics Paschalis Korosoglou (support@grid.auth.gr) User and Application Support Unit Scientific Computing Center @ AUTH Overview of Parallel computing
More informationMPI Program Structure
MPI Program Structure Handles MPI communicator MPI_COMM_WORLD Header files MPI function format Initializing MPI Communicator size Process rank Exiting MPI 1 Handles MPI controls its own internal data structures
More informationIntroduction to MPI. Ricardo Fonseca. https://sites.google.com/view/rafonseca2017/
Introduction to MPI Ricardo Fonseca https://sites.google.com/view/rafonseca2017/ Outline Distributed Memory Programming (MPI) Message Passing Model Initializing and terminating programs Point to point
More informationMessage-Passing Thought Exercise
Message-Passing Thought Exercise Traffic Modelling David Henty EPCC d.henty@epcc.ed.ac.uk +44 131 650 5960 traffic flow 15 May 2015 Traffic Modelling 2 we want to predict traffic flow to look for effects
More informationHigh Performance Computing Course Notes Message Passing Programming III
High Performance Computing Course Notes 2008-2009 2009 Message Passing Programming III Communication modes Synchronous mode The communication is considered complete when the sender receives the acknowledgement
More informationPreventing and Finding Bugs in Parallel Programs
Preventing and Finding Bugs in Parallel Programs Charlie Peck Earlham College BW/UPEP Institute @ NCSA May, 2011 1 How Did We Get Here? Debugging serial programs can be hard; debugging parallel programs
More informationL14 Supercomputing - Part 2
Geophysical Computing L14-1 L14 Supercomputing - Part 2 1. MPI Code Structure Writing parallel code can be done in either C or Fortran. The Message Passing Interface (MPI) is just a set of subroutines
More informationMATH 676. Finite element methods in scientific computing
MATH 676 Finite element methods in scientific computing Wolfgang Bangerth, Texas A&M University Lecture 41.25: Parallelization on a cluster of distributed memory machines Part 2: Debugging with MPI Debugging
More informationThe Message Passing Model
Introduction to MPI The Message Passing Model Applications that do not share a global address space need a Message Passing Framework. An application passes messages among processes in order to perform
More informationAMath 483/583 Lecture 18 May 6, 2011
AMath 483/583 Lecture 18 May 6, 2011 Today: MPI concepts Communicators, broadcast, reduce Next week: MPI send and receive Iterative methods Read: Class notes and references $CLASSHG/codes/mpi MPI Message
More informationA few words about MPI (Message Passing Interface) T. Edwald 10 June 2008
A few words about MPI (Message Passing Interface) T. Edwald 10 June 2008 1 Overview Introduction and very short historical review MPI - as simple as it comes Communications Process Topologies (I have no
More informationHigh Performance Computing Course Notes Message Passing Programming III
High Performance Computing Course Notes 2009-2010 2010 Message Passing Programming III Blocking synchronous send the sender doesn t return until it receives the acknowledgement from the receiver that the
More informationIntroduction to MPI. Ekpe Okorafor. School of Parallel Programming & Parallel Architecture for HPC ICTP October, 2014
Introduction to MPI Ekpe Okorafor School of Parallel Programming & Parallel Architecture for HPC ICTP October, 2014 Topics Introduction MPI Model and Basic Calls MPI Communication Summary 2 Topics Introduction
More informationParallel Programming, MPI Lecture 2
Parallel Programming, MPI Lecture 2 Ehsan Nedaaee Oskoee 1 1 Department of Physics IASBS IPM Grid and HPC workshop IV, 2011 Outline 1 Introduction and Review The Von Neumann Computer Kinds of Parallel
More informationSupercomputing in Plain English Exercise #6: MPI Point to Point
Supercomputing in Plain English Exercise #6: MPI Point to Point In this exercise, we ll use the same conventions and commands as in Exercises #1, #2, #3, #4 and #5. You should refer back to the Exercise
More informationCINES MPI. Johanne Charpentier & Gabriel Hautreux
Training @ CINES MPI Johanne Charpentier & Gabriel Hautreux charpentier@cines.fr hautreux@cines.fr Clusters Architecture OpenMP MPI Hybrid MPI+OpenMP MPI Message Passing Interface 1. Introduction 2. MPI
More informationFROM SERIAL FORTRAN TO MPI PARALLEL FORTRAN AT THE IAS: SOME ILLUSTRATIVE EXAMPLES
FROM SERIAL FORTRAN TO MPI PARALLEL FORTRAN AT THE IAS: SOME ILLUSTRATIVE EXAMPLES Let's begin with a simple trapezoidal integration program. This version creates 'nbin' bins, and sums the function `fcn'
More informationProgramming with MPI
Programming with MPI p. 1/?? Programming with MPI Attributes and I/O Nick Maclaren nmm1@cam.ac.uk May 2008 Programming with MPI p. 2/?? Address-Sized Integers MPI needs address--sized integers Mostly for
More informationElementary Parallel Programming with Examples. Reinhold Bader (LRZ) Georg Hager (RRZE)
Elementary Parallel Programming with Examples Reinhold Bader (LRZ) Georg Hager (RRZE) Two Paradigms for Parallel Programming Hardware Designs Distributed Memory M Message Passing explicit programming required
More informationcharacter :: buffer(100) integer :: position real :: a, b integer :: n position = 0 call MPI_PACK(a, 1, MPI_REAL, buffer, 100, & position, MPI_COMM_WO
MPI_PACK and MPI_UNPACK Each communication incurs a latency penalty so it is best to group communications together Requires data to be contiguous in memory with no gaps between variables This is true for
More informationCS 426. Building and Running a Parallel Application
CS 426 Building and Running a Parallel Application 1 Task/Channel Model Design Efficient Parallel Programs (or Algorithms) Mainly for distributed memory systems (e.g. Clusters) Break Parallel Computations
More informationIntroduction to the Message Passing Interface (MPI)
Introduction to the Message Passing Interface (MPI) CPS343 Parallel and High Performance Computing Spring 2018 CPS343 (Parallel and HPC) Introduction to the Message Passing Interface (MPI) Spring 2018
More informationAn Introduction to MPI
An Introduction to MPI Parallel Programming with the Message Passing Interface William Gropp Ewing Lusk Argonne National Laboratory 1 Outline Background The message-passing model Origins of MPI and current
More informationParallelization. Tianhe-1A, 2.45 Pentaflops/s, 224 Terabytes RAM. Nigel Mitchell
Parallelization Tianhe-1A, 2.45 Pentaflops/s, 224 Terabytes RAM Nigel Mitchell Outline Pros and Cons of parallelization Shared memory vs. cluster computing MPI as a tool for sending and receiving messages
More informationParallel Programming using MPI. Supercomputing group CINECA
Parallel Programming using MPI Supercomputing group CINECA Contents Programming with message passing Introduction to message passing and MPI Basic MPI programs MPI Communicators Send and Receive function
More informationMPI Lab. Steve Lantz Susan Mehringer. Parallel Computing on Ranger and Longhorn May 16, 2012
MPI Lab Steve Lantz Susan Mehringer Parallel Computing on Ranger and Longhorn May 16, 2012 1 MPI Lab Parallelization (Calculating p in parallel) How to split a problem across multiple processors Broadcasting
More informationImplementation of Parallelization
Implementation of Parallelization OpenMP, PThreads and MPI Jascha Schewtschenko Institute of Cosmology and Gravitation, University of Portsmouth May 9, 2018 JAS (ICG, Portsmouth) Implementation of Parallelization
More informationLaplace Exercise Solution Review
Laplace Exercise Solution Review John Urbanic Parallel Computing Scientist Pittsburgh Supercomputing Center Copyright 2018 Finished? If you have finished, we can review a few principles that you have inevitably
More informationPractical Introduction to Message-Passing Interface (MPI)
1 Outline of the workshop 2 Practical Introduction to Message-Passing Interface (MPI) Bart Oldeman, Calcul Québec McGill HPC Bart.Oldeman@mcgill.ca Theoretical / practical introduction Parallelizing your
More informationIntroduction to Parallel Programming with MPI
Introduction to Parallel Programming with MPI PICASso Tutorial October 25-26, 2006 Stéphane Ethier (ethier@pppl.gov) Computational Plasma Physics Group Princeton Plasma Physics Lab Why Parallel Computing?
More information30 Nov Dec Advanced School in High Performance and GRID Computing Concepts and Applications, ICTP, Trieste, Italy
Advanced School in High Performance and GRID Computing Concepts and Applications, ICTP, Trieste, Italy Why serial is not enough Computing architectures Parallel paradigms Message Passing Interface How
More informationExercises: Message-Passing Programming
T E H U I V E R S I T Y O H F R G E I B U Exercises: Message-Passing Programming Hello World avid Henty. Write an MPI program which prints the message Hello World.. Compile and run on several processes
More informationMessage-Passing and MPI Programming
Message-Passing and MPI Programming More on Collectives N.M. Maclaren Computing Service nmm1@cam.ac.uk ext. 34761 July 2010 5.1 Introduction There are two important facilities we have not covered yet;
More informationLaplace Exercise Solution Review
Laplace Exercise Solution Review John Urbanic Parallel Computing Scientist Pittsburgh Supercomputing Center Copyright 2017 Finished? If you have finished, we can review a few principles that you have inevitably
More informationIntroduction to MPI Part II Collective Communications and communicators
Introduction to MPI Part II Collective Communications and communicators Andrew Emerson, Fabio Affinito {a.emerson,f.affinito}@cineca.it SuperComputing Applications and Innovation Department Collective
More informationIntroduction to MPI. SHARCNET MPI Lecture Series: Part I of II. Paul Preney, OCT, M.Sc., B.Ed., B.Sc.
Introduction to MPI SHARCNET MPI Lecture Series: Part I of II Paul Preney, OCT, M.Sc., B.Ed., B.Sc. preney@sharcnet.ca School of Computer Science University of Windsor Windsor, Ontario, Canada Copyright
More informationMessage-Passing Programming with MPI
Message-Passing Programming with MPI Message-Passing Concepts David Henty d.henty@epcc.ed.ac.uk EPCC, University of Edinburgh Overview This lecture will cover message passing model SPMD communication modes
More informationWelcome. HRSK Practical on Debugging, Zellescher Weg 12 Willers-Bau A106 Tel
Center for Information Services and High Performance Computing (ZIH) Welcome HRSK Practical on Debugging, 03.04.2009 Zellescher Weg 12 Willers-Bau A106 Tel. +49 351-463 - 31945 Matthias Lieber (matthias.lieber@tu-dresden.de)
More informationDistributed Memory Programming With MPI (4)
Distributed Memory Programming With MPI (4) 2014 Spring Jinkyu Jeong (jinkyu@skku.edu) 1 Roadmap Hello World in MPI program Basic APIs of MPI Example program The Trapezoidal Rule in MPI. Collective communication.
More informationUSER-DEFINED DATATYPES
Advanced MPI USER-DEFINED DATATYPES MPI datatypes MPI datatypes are used for communication purposes Datatype tells MPI where to take the data when sending or where to put data when receiving Elementary
More informationFibonacci in Lisp. Computer Programming: Skills & Concepts (CP1) Programming Languages. Varieties of Programing Language
Fibonacci in Lisp Computer Programming: Skills & Concepts (CP1) Programming Languages (defun fibonacci (n) (if (or (= n 0) (= n 1)) 1 (+ (fibonacci (- n 1)) (fibonacci (- n 2))))) 22nd November 2010 defun-
More informationlslogin3$ cd lslogin3$ tar -xvf ~train00/mpibasic_lab.tar cd mpibasic_lab/pi cd mpibasic_lab/decomp1d
MPI Lab Getting Started Login to ranger.tacc.utexas.edu Untar the lab source code lslogin3$ cd lslogin3$ tar -xvf ~train00/mpibasic_lab.tar Part 1: Getting Started with simple parallel coding hello mpi-world
More informationIPM Workshop on High Performance Computing (HPC08) IPM School of Physics Workshop on High Perfomance Computing/HPC08
IPM School of Physics Workshop on High Perfomance Computing/HPC08 16-21 February 2008 MPI tutorial Luca Heltai Stefano Cozzini Democritos/INFM + SISSA 1 When
More informationProgramming with MPI
Programming with MPI p. 1/?? Programming with MPI Miscellaneous Guidelines Nick Maclaren nmm1@cam.ac.uk March 2010 Programming with MPI p. 2/?? Summary This is a miscellaneous set of practical points Over--simplifies
More informationIntroduction to Parallel Programming. Martin Čuma Center for High Performance Computing University of Utah
Introduction to Parallel Programming Martin Čuma Center for High Performance Computing University of Utah mcuma@chpc.utah.edu Overview Types of parallel computers. Parallel programming options. How to
More informationCEE 618 Scientific Parallel Computing (Lecture 5): Message-Passing Interface (MPI) advanced
1 / 32 CEE 618 Scientific Parallel Computing (Lecture 5): Message-Passing Interface (MPI) advanced Albert S. Kim Department of Civil and Environmental Engineering University of Hawai i at Manoa 2540 Dole
More informationPaul Burton April 2015 An Introduction to MPI Programming
Paul Burton April 2015 Topics Introduction Initialising MPI & basic concepts Compiling and running a parallel program on the Cray Practical : Hello World MPI program Synchronisation Practical Data types
More informationOptimization of MPI Applications Rolf Rabenseifner
Optimization of MPI Applications Rolf Rabenseifner University of Stuttgart High-Performance Computing-Center Stuttgart (HLRS) www.hlrs.de Optimization of MPI Applications Slide 1 Optimization and Standardization
More informationIntroduction to MPI, the Message Passing Library
Chapter 3, p. 1/57 Basics of Basic Messages -To-? Introduction to, the Message Passing Library School of Engineering Sciences Computations for Large-Scale Problems I Chapter 3, p. 2/57 Outline Basics of
More informationProgramming with MPI
Programming with MPI p. 1/?? Programming with MPI More on Datatypes and Collectives Nick Maclaren nmm1@cam.ac.uk May 2008 Programming with MPI p. 2/?? Less Basic Collective Use A few important facilities
More informationParallel Paradigms & Programming Models. Lectured by: Pham Tran Vu Prepared by: Thoai Nam
Parallel Paradigms & Programming Models Lectured by: Pham Tran Vu Prepared by: Thoai Nam Outline Parallel programming paradigms Programmability issues Parallel programming models Implicit parallelism Explicit
More informationIntroduction to Parallel Programming. Martin Čuma Center for High Performance Computing University of Utah
Introduction to Parallel Programming Martin Čuma Center for High Performance Computing University of Utah m.cuma@utah.edu Overview Types of parallel computers. Parallel programming options. How to write
More informationMessage-Passing and MPI Programming
Message-Passing and MPI Programming 5.1 Introduction More on Datatypes and Collectives N.M. Maclaren nmm1@cam.ac.uk July 2010 There are a few important facilities we have not covered yet; they are less
More informationMPI introduction - exercises -
MPI introduction - exercises - Paolo Ramieri, Maurizio Cremonesi May 2016 Startup notes Access the server and go on scratch partition: ssh a08tra49@login.galileo.cineca.it cd $CINECA_SCRATCH Create a job
More informationMPI MESSAGE PASSING INTERFACE
MPI MESSAGE PASSING INTERFACE David COLIGNON CÉCI - Consortium des Équipements de Calcul Intensif http://hpc.montefiore.ulg.ac.be Outline Introduction From serial source code to parallel execution MPI
More informationHPC Architectures. Types of resource currently in use
HPC Architectures Types of resource currently in use Reusing this material This work is licensed under a Creative Commons Attribution- NonCommercial-ShareAlike 4.0 International License. http://creativecommons.org/licenses/by-nc-sa/4.0/deed.en_us
More informationPeter Pacheco. Chapter 3. Distributed Memory Programming with MPI. Copyright 2010, Elsevier Inc. All rights Reserved
An Introduction to Parallel Programming Peter Pacheco Chapter 3 Distributed Memory Programming with MPI 1 Roadmap Writing your first MPI program. Using the common MPI functions. The Trapezoidal Rule in
More informationMPI Lab. How to split a problem across multiple processors Broadcasting input to other nodes Using MPI_Reduce to accumulate partial sums
MPI Lab Parallelization (Calculating π in parallel) How to split a problem across multiple processors Broadcasting input to other nodes Using MPI_Reduce to accumulate partial sums Sharing Data Across Processors
More informationDistributed Memory Programming with MPI. Copyright 2010, Elsevier Inc. All rights Reserved
An Introduction to Parallel Programming Peter Pacheco Chapter 3 Distributed Memory Programming with MPI 1 Roadmap Writing your first MPI program. Using the common MPI functions. The Trapezoidal Rule in
More informationLecture V: Introduction to parallel programming with Fortran coarrays
Lecture V: Introduction to parallel programming with Fortran coarrays What is parallel computing? Serial computing Single processing unit (core) is used for solving a problem One task processed at a time
More informationIntroduction to MPI part II. Fabio AFFINITO
Introduction to MPI part II Fabio AFFINITO (f.affinito@cineca.it) Collective communications Communications involving a group of processes. They are called by all the ranks involved in a communicator (or
More information